Processes/Strategies

This task characteristic refers to the way in which examinees process text to respond correctly to a question or directive. It includes the processes used to relate information in the question (the given information) to the necessary information in the text (the new information), as well as the processes needed to either identify or construct the correct response from the information available. Three variables in the reading/literacy research used to investigate tasks from national and international surveys will be considered here. These are: type of match, type of information requested, and plausibility of distracting information. They are briefly described here. They are characterized through a discussion of exemplary tasks in the next section and fully operationalized in the appendix at the end of this paper.

Type of match

Four types of matching strategies were identified: locating, cycling, integrating, and generating. Locating tasks require examinees to match one or more features of information stated in the question to either identical or synonymous information provided in the text. Cycling tasks also require examinees to match one or more features of information, but unlike locating tasks, they require respondents to engage in a series of feature matches to satisfy conditions stated in the question. Integrating tasks require examinees to pull together two or more pieces of information from the text according to some type of specified relationship. For example, this relationship might call for examinees to identify similarities (i.e., make a comparison), differences (i.e., contrast), degree (i.e., smaller or larger), or cause-and-effect relationships. This information may be located within a single paragraph or it may appear in different paragraphs or sections of the text. In integrating information, examinees draw upon information categories provided in a question to locate the corresponding information in the text. They then relate the text information associated with these different categories based upon the relationship term specified in the question. In some cases, however, examinees must generate these categories and/or relationships before integrating the information stated in the text.

In addition to requiring examinees to apply one of these four strategies, the type of match between a question and the text is influenced by several other processing conditions that contribute to a task's overall difficulty. The first of these is the number of phrases that must be used in the search. Task difficulty increases with the amount of information in the question for which the examinee must search in the text. For instance, questions that consist of only one independent clause tend to be easier, on average, than those that contain several independent or dependent clauses. Difficulty also increases with the number of responses that examinees are asked to provide. Questions that request a single answer are easier than those that require three or more answers. Further, questions that specify the number of responses tend to be easier than those that do not. For example, a question that states, "List the three reasons . . ." would be easier than one that said, "List the reasons . . ." Tasks are also influenced by the degree to which examinees have to make inferences to match the given information in a question to corresponding information in the text, and to identify the requested information. An additive scoring model defining type of match for prose and document literacy tasks is provided in Appendix A.